Epistatic interactions between thiopurine methyltransferase (TPMT ...

Eur J Clin Pharmacol (2012) 68:379–387 DOI 10.1007/s00228-011-1133-1

PHARMACOGENETICS

Epistatic interactions between thiopurine methyltransferase (TPMT) and inosine triphosphate pyrophosphatase (ITPA) variations determine 6-mercaptopurine toxicity in Indian children with acute lymphoblastic leukemia Patchva Dorababu & Narayana Nagesh & Vijay Gandhi Linga & Sadashivudu Gundeti & Vijay Kumar Kutala & Pallu Reddanna & Raghunadharao Digumarti

Received: 3 June 2011 / Accepted: 28 September 2011 / Published online: 19 October 2011 # Springer-Verlag 2011

Abstract Purpose To explore the role of genetic variants of thiopurine methyltransferase (TPMT) and inosine triphosphate pyrophosphatase (ITPA) in 6-mercaptopurine (6-MP)-induced toxicity in Indian children with acute lymphoblastic leukemia (ALL). Methods Children with ALL receiving 6-MP in maintenance phase of treatment (n=90) were enrolled in the study. Bidirectional sequencing of TPMT (whole gene) and ITPA (exon 2, exon 3, and intron 2) was undertaken, and correlation between genotype and 6-MP toxicity was assessed. Results Five variations were observed in TPMT, including two exonic variations, TPMT*12 (374 C>T) and TPMT*3C (719A>G), and three intronic, intron 3 (12356 C>T), intron P. Dorababu : V. G. Linga : S. Gundeti : R. Digumarti (*) Department of Medical Oncology, Nizam’s Institute of Medical Sciences, Panjagutta, Hyderabad, Andhra Pradesh, India PIN: 500082 e-mail: [email protected] N. Nagesh Centre for Cellular and Molecular Biology, Hyderabad, India V. K. Kutala Department of Clinical Pharmacology and Therapeutics, Nizam’s Institute of Medical Sciences, Hyderabad, India P. Reddanna Department of Animal Sciences, University of Hyderabad, Hyderabad, India

4 (16638 C>T), and TPMT rs2842949. Two exonic, ITPA exon −2 (94 C→A) and exon 3 of ITPA (138 G>A), and one intronic, ITPA intron 2 (A→C), variations were observed in ITPA. Multifactor dimensionality reduction analysis of all the genetic variants showed independent association of ITPA 94 C→A as well as synergic epistatic interactions, i.e., TPMT*12 × ITPA ex3, ITPA ex2 × TPMT*12 × ITPA ex3, and TPMT*3C × ITPA ex2 × TPMT*12 × ITPA ex3, in determining hematological toxicity. This is further substantiated by a multiple linear regression model, which showed moderate predictability of toxicity with these variants (area under the curve=0.70, p=0.004). Conclusion Our results suggest that apart from the individual effect of ITPA 94 C→A, epistatic interactions between the variations of TPMT (*3C, *12) and ITPA (ex2, ex3) are associated with the 6-MP toxicity. Testing these variants facilitates tailoring of the 6-MP therapy in children with ALL. Keywords Pharmacogenetics . Thiopurine methyltransferase . Inosine triphosphate pyrophosphatase . 6-Mercaptopurine . Childhood acute lymphoblastic leukemia

Introduction 6-Mercaptopurine (6-MP) is a widely used drug during the maintenance treatment of childhood acute lymphoblastic leukemia (ALL), the most common childhood cancer. Following oral administration, 6-MP undergoes extensive metabolism by three competing enzymatic pathways catalyzed by xanthine oxidase (XO), thiopurine S-methyltransferase (TPMT), and

380

hypoxanthine guanine phosphoribosyl transferase (HGPRT). The TPMT pathway plays a pivotal role in intracellular inactivation of 6-MP in hematopoietic tissue by converting methyl thioinosine monophosphate (me-TIMP) and 6-methyl mercaptopurine (6-MMP) [1, 2]. HGPRT metabolizes 6MP to 6-thioinosine monophosphate (TIMP). It is then subsequently converted to nucleotides of 6-thioguanine (6TGN) through a two-step process involving inosine monophosphate dehydrogenase (IMPDH) and guanosine monophosphate synthase (GMPS) [3]. Inosine triphosphate pyrophosphohydrolase (ITPA) is another enzyme involved in this pathway, which catalyzes the pyrophosphohydrolysis of inosine triphosphate (ITP) to inosine monophosphate (IMP) and prevents the accumulation of potentially toxic compounds such as inosine triphosphate (ITP) and deoxy inosine triphosphate (de-ITP). In ITPA-deficient patients, the accumulation of toxic compounds, i.e., ITP and de-ITP, was demonstrated; however, their association with any pathology is still debated [4]. Incorporation of deoxy-6-thioguanosine5-triphosphate (dTGTP) into DNA triggers cell cycle arrest and apoptosis by a process involving the mismatch repair (MMR). The principal mechanism of action of 6-MP is considered to be due to 6-TGTP nucleotide-induced lymphocyte apoptosis [5] (Schema 1). The genetic polymorphisms in TPMT may influence the variability of balance between 6-TGN and 6-MMPN. The association between the polymorphisms in TPMT and 6-mercaptopurine toxicity is one of the best-studied examples in pharmacogenetics [6, 7]. TPMT activity

Schema 1 6-Mercaptopurine metabolic pathway. TPMT Thiopurine methyltransferase, XO xanthine oxidase, HGPRT hypoxanthineguanine-phosphoribosyl transferase, IMPDH inosine monophosphate dehydrogenase, GMPS guanosine monophosphate synthatase, ITPA inosine triphosphate pyrophosphatase, 6-MMP methyl mercaptopurine, 6-TUA 6-thiouric acid, TIMP 6-thioinosine-5′- monophosphate, 6TXMP 6-thioxanthine -5′- monophosphate, 6-TGMP 6-thioguanosine5′- monophosphate, 6-TGNs 6-thioguanine nucleotides

Eur J Clin Pharmacol (2012) 68:379–387

showed inverse correlation to TGN concentrations in the erythrocytes of children treated for leukemia [8]. High erythrocyte concentrations of TGNs were associated with the degree of leucopenia and a good prognosis [9], whereas low concentrations of TGN or high TPMT activity showed association with an increased risk for relapse and increased incidence of infectious diseases, respectively [10, 11]. The TPMT protein is encoded by a 27-kb gene on human chromosome 6p22.3. The TPMT gene has ten exons, eight of which encode the 28-kDa protein. To date, at least 30 variant alleles of TPMT have been identified, and the functionally relevant variants that lower TPMT activity in the majority of individuals are TPMT*2, TPMT*3A, TPMT*3B, and TPMT*3C. All four of these TPMT variants are the most common variants observed in three major populations, i.e., Caucasians, Asians, and Africans [12–15]. The association of these TPMT variants with TPMT phenotype is well established [12]. Approximately 0.3% of patients are homozygous for a variant allele and have zero or very low enzyme activity, while 5–15% of patients are heterozygous and have intermediate enzymatic activity [14, 15]. However, lower TPMT enzyme activity or the presence of a variant allele has been observed to increase the risk of thiopurinerelated drug toxicity, particularly with 6-MP [16–19]. This emphasizes the need for either molecular analysis of the TPMT gene or measurement of TPMT activity in patients with ALL treated with 6-MP to reduce the risk of drugrelated toxicity. Patients with undetectable TPMT activity might benefit from alternative therapeutic strategies. The ITPA activity is also genetically determined. Patients homozygous for a 94 C>A (Pro32Thr, rs1127354) variant are characterized by zero or low enzyme activity, whereas homozygotes for the other functional polymorphisms in intron 2 (IVS2 + 21A>C, rs7270101) and compound heterozygous subjects express∼30% of the wild-type activity [20, 21]. ITPA has been studied in the context of mercaptopurine toxicity in several diseases, including inflammatory bowel disease (IBD), ALL, and in patients with transplant rejection [22]. Since thiopurine metabolism is regulated by multiple enzymes, testing only TPMT genotype may not give sufficient information about the hematological toxicity. Studies focused on the multi-locus investigation will help in better understanding 6-MP-mediated toxicity in ALL patients. Previous studies from India focused on the analysis of TPMT polymorphism using conventional approaches such as PCR-RFLP and ASO-PCR. However, there are no studies reported in the literature where complete gene analysis of two enzymes was done in the Indian population and correlated with 6-MP tolerance. Hence, the aim of the study was to screen the complete gene of TPMT and the most common variants in ITPA, i.e., exon 2, exon 3, and intron 2, and to establish an association between genotype and 6-MP toxicity in patients with ALL.


381

Materials and methods Patients A total of 90 ALL children who are on maintenance dose of 6-MP were recruited in the study from the outpatient unit of Medical Oncology, Nizam’s Institute of Medical Sciences, Hyderabad, during the period of June 2009 to April 2011. This study was approved by the Institutional Ethical Committee of Nizam’s Institute of Medical Sciences (NIMS), Hyderabad, India (EC/ NIMS/1123/2009). Informed consent was obtained from each subject’s caregiver. All patients enrolled in the study were treated at our institution according to MCP-841 Protocol [23]. During the maintenance phase of ALL treatment, an oral dose of 75 mg/m2 of 6-MP was given daily for 3 weeks, skipping every fourth week, for a total duration of 12 weeks. On days 43, 71, and 99 of the maintenance phase, a complete hemogram was done, and the average decrease in total

Table 1 PCR conditions for genetic analysis

leukocyte count during these three occasions was taken as a measure of toxicity. If any patient had drug-related toxicity, then the dose of 6-MP was reduced by 10– 20%. The toxicity grading was done using the National Cancer Institute Common Toxicity Criteria (NCI CTC) version 3. During the 6-MP therapy, a dose of methotrexate (MTX), 15 mg/m2 PO, was given once a week for 3 weeks. The doses of both drugs were “titrated” to keep the white blood cell (WBC) count in the range of 2,000–3,000 mm−3. DNA isolation and PCR amplification Blood samples were collected in 5 ml EDTA tubes. Genomic DNA was extracted from whole blood by the standard phenol-chloroform method. The entire TPMT gene and exon 2, exon 3, and intron 2 of ITPA gene fragments were amplified using oligonucleotide primers and conditions as shown in Table 1. The reason for choosing only two exons and one intron in the ITPA gene

Exon

Forward primer (5′ to 3′)

Reverse primer (5′ to 3′)

Product length (bp)

PCR conditions

TPMT 2

ataggttttcatttagttcatcaa

catttttgtgtctcccgagt

394

D:95°C/30 s A:60°C/30 s E:72°C/1 min (35×)

TPMT 3

TPMT 4

TPMT 5

D Denaturation, A annealing, E extension, × number of cycles

aaagccctgggtgtaagtca

tgaaaccctatgaacctgaatt

ccctagaacttttgctttgct

caaaactcaatccagaaagact

gtgaatctgcgtgctaaatag

tgcctcagtttcccatagttt

508

D:95°C/30 s

380

A:50°C/30 s E:72°C/1 min (35×) D:95°C/30 s

369

A:60°C/30 s E:72°C/1 min (35×) D:95°C/30 s

TPMT 6

cgcagacgtgagatcctaat

agccacaagccttatagcct

411

TPMT 7

ctcagtagtatcagcgaaagta

aagaaactaggcaactggtaaaa

245

TPMT 8

tgccatcctctcagtaagtca

cagcacgccaggcccaaaa

281

TPMT10

gttgggattacaggtgtgag

agtttagttacatctttttcttct

420

ITPA 2&3

atgagaaaggcggatgacag

tgttgagctttctaggaccg

403

A:60°C/30 s E:72°C/1 min D:95°C/30 s A:60°C/30 s E:72°C/1 min D:95°C/30 s A:60°C/30 s E:72°C/1 min D:95°C/30 s A:60°C/30 s E:72°C/1 min

(35×)

(35×)

(35×)

(35×)

D:94°C/30 s A:58°C/30 s E:72°C/1 min (33×) D:95°C/30 s A:58°C/30 s E:72°C/2 min (35×)

382

was to focus on the variants that are commonly associated with hematological toxicity in other ethnic groups and populations. PCR was performed in a volume of 10 μl containing 50 ng of genomic DNA, 10 pmol of each primer, and 5 μl of PCR Master Mix (2×) (Fermentas Life Sciences). Sequencing analysis The sequencing reactions were performed bidirectionally from at least two independent amplification products. After dye termination PCR, amplicons were purified using isopropanol precipitation, diluted, and cycle-sequenced using the ABI Big Dye Terminator kit v3.1 (ABI, Foster City, CA, USA) according to the manufacturer’s instructions. Sequencing reactions were electrophoresed on an ABI3100 genetic analyzer. Electrophoregrams were analyzed in both sense and antisense directions for the presence of variants (Fig. 1). To rule out the genotyping errors, 10% of the specimens were recollected and reanalyzed by another investigator, which rules out pre-analytical and post-analytical errors. Using this quality control approach, we obtained 100% concordance in the results. Statistical analysis All the genotyping data were computed as 0, 1, and 2 based on the number of variant alleles, thus representing wild type, heterozygous, and homozygous mutant genotypes. The deviation from Hardy-Weinberg equilibrium (HWE) was tested using the chi-squared test. Hematological toxicity grading was correlated with different allelic variants using Fig. 1 Electrophorogram of high-risk variants of TPMT and ITPA. a TPMT*3C (719 A→G) wild type, b TPMT*3C (719 A→G) heterozygous, c TPMT rs2842949 (G→T) wild type, d TPMT rs2842949 (G→T) homozygous mutant, e ITPA Ex2 (94 C→A) wild type, f ITPA Ex2 (94 C→A) heterozygous


Spearman Rank Correlation Coefficient (r). Gene-dosage effect on total leukocyte count reduction was evaluated by analysis of variance (ANOVA) across a number of highrisk alleles. A p value of A variations after 6-MP treatment. Total leukocyte count and platelet count were reduced in patients harboring the wild-type allele at TPMT and ITPA loci. Variant alleles at TPMT and ITPA showed evidence of


383

Table 2 Association of genetic variants with clinical and hematological parameters in ALL patients treated with 6-MP Wild-type genotype at three loci

TMPT*3C

ITPA 94 C>A

TPMT rs2842949

Pre-hemoglobin (g/dl) Post-hemoglobin (g/dl)

10.69±1.61 10.28±1.78

10.41±0.10 9.68±2.13

9.88±1.34 9.18±2.72

11.1±0.93 10.55±2.96

P value Pre-total leukocyte count (cells/mm3) Post-total leukocyte count (cells/mm3) P value Pre-platelet count (lakhs/mm3) Post-platelet count (lakhs/mm3)

0.18 7,464±4,805 3,530±2,022

0.54 10,800±8,705 1,337±879

0.66 5,410±2,756 1,320±630

0.72 6,950±2,043 1,325±427

0.0001 2.49±1.59 1.90±1.23

0.12 2.03±0.81 0.79±0.74

0.04 3.03±1.08 1.33±0.81

0.008 2.58±1.17 1.73±0.32

P value Febrile episodes (% of subjects)

0.005 49.18%

0.12 100 %

0.10 80%

0.22 25%

Dose reductions (% of subjects)

31.14%

100%

25%

50%

Relapse/death Dose of 6-MP in mg m−2 day−1

8.3% 46.19±20.97

25% 21.88±6.25

0% 41.67±12.91

0% 46.88±32.87

SD Standard deviation, MP mercaptopurine

leucopenia post-treatment. TPMT*3C and ITPA 94 C>A variants showed high frequency of febrile episodes compared to wild-type genotype (at TPMT/ITPA loci) and other variants. All the patients harboring TPMT*3C were adjusted to lower dose based on the hematological toxicity profile. Among the TPMT*3C variants, 25% showed either relapse or death. Genotyping of TPMT and ITPA was done in all patients. Allelic variants of TPMT and ITPA genes and their frequencies are shown in Table 3. Two exonic and three intronic variations were observed in TPMT. The minor allele frequencies of TPMT*12 (374 C>T), TPMT*3C (719A>G), intron 3 (12356 C>T), intron 4 (16638 C>T), and TPMT rs2842949 (intron 7, 26354 G>T) were 30.6, 2.2, 8.9, 2.8, and 3.9%, respectively. As shown in Table 3, two exonic variations were observed in the ITPA gene, i.e., Table 3 Frequency of TPMT and ITPA alleles

WW Wild, WM heterozygous, MM homozygous mutant, W number of wild alleles, M number of mutant alleles, MAF minor allele frequency, hwe Hardy-Weinberg equilibrium

ITPA exon 2 (94 C→A) and exon 3 of ITPA (138 G>A), and one intronic variation, i.e., ITPA Int2 (A→C), with variant allele frequencies of 5.0, 25.6, and 0.6%, respectively. Around 50% of variants showed distribution in accordance with Hardy-Weinberg equilibrium (Table 3). The variations that are associated with hematological toxicity in univariate analysis were rechecked for Hardy-Weinberg equilibrium using control samples, which were in accordance with the equilibrium (Table 3). The toxicity grading was done using NCI CTC criteria and was correlated with the different genotypes using Spearman Rank Correlation Coefficient. The results are shown in Table 4. Univariate analysis showed association of TPMT*3C, TPMT rs2842949, and ITPA (94 C>A) with leucopenia (Table 4). However, MDR analysis revealed independent effects of ITPA ex2 polymorphism as well as

Variation

rs Number

WW

WM

MM

M/W

MAF

p (hwe)

Patients TPMT*3C (A→G) TPMT*12 (C→T) TPMT Int7 (G→T)

rs1142345 rs2842934 rs2842949

85 52 86

4 21 1

0 17 3

4/174 55/125 7/173

0.022 0.306 0.039

1.00 0.0001 0.016

rs4449636 rs2518463 rs1127354 rs8362 rs7270101

78 85 81 61 89

8 5 9 12 1

4 0 0 17 0

16/164 5/175 9/171 46/134 1/179

0.089 0.028 0.050 0.256 0.006

0.003 1.00 1.00 0.0001 1.00

rs1142345 rs2842949 rs1127354

97 59 87

3 32 13

0 9 0

3/197 50/150 13/187

0.015 0.25 0.065

1.00 0.27 1.00

TPMT Int.3 (C→T) TPMT Int.4 (T→C) ITPA Exon-2 (C→A) ITPA Exon-3 (G→A) ITPA Int.2 (A→C) Controls TPMT*3C (A→G) TPMT Int7 (G→T) ITPA EX2 (C→A)

384


Table 4 Correlation between the genetic variants and hematological toxicity Polymorphism

rs Number

Correlation coefficient (r)

p Value

TPMT*3C

rs1142345

0.28

0.015*

TPMT*12

rs2842934

0.21

0.07

TPMT Int7 TPMT Int.3

rs2842949 rs4449636

0.28 −0.13

0.015* 0.27

TPMT Int.4

rs2518463

0.077

0.51

ITPA Exon-2 ITPA Exon-3

rs1127354 rs8362

0.30 0.093

0.009* 0.43

ITPA Int.2

rs7270101

0.033

0.78

*p < 0.05

synergic gene-gene (epistatic) interactions, i.e., TPMT*12 × ITPA ex3, ITPA ex2 × TPMT*12 × ITPA ex3, and TPMT*3C × ITPA ex2 × TPMT*12 × ITPA ex3 in determining hematological toxicity (Table 5). As shown in Fig. 2, there was a dose-dependent reduction in the total leukocyte count with an increase in the number of high-risk alleles of TPMT and ITPA loci (as observed in interaction model), thus explaining the higher toxicity with an increase in the number of risk alleles. The multiple linear regression model was developed to predict hematological toxicity based on TPMT and ITPA genotypes and showed moderate predictability of toxicity with area under the curve of 0.70 (p=0.004) (Fig. 3). TPMT and ITPA genotypes together can explain 27% of the variability in hematological toxicity.

Discussion The use of human genome sequence information for delivering safe and effective medication by developing “personalized medicine” is bridging the gap between the basic research and clinical practice. The current study gains importance as there are no systematic molecular analyses on Indian subjects to identify the pharmacogenomic determinants of 6-MP-associated drug toxicity. A whole-

gene sequencing of TPMT and the most common functional variants of ITPA helped in establishing the variations that are prevalent in Indian subjects and their association with hematological toxicity. This study reconfirms the existing evidence that impaired 6-MP metabolism due to genetic variations in TPMT and ITPA induces drug toxicity. Although the importance of the TPMT genetic polymorphism is very well known and characterized [1, 24–26], this is the first report from India showing a synergetic effect of TPMT and ITPA genotypes in optimizing 6-MP therapy based on genetic constitution. Previous studies have shown that TPMT*3C is the most common variant in Indians with its frequency ranging between 0.8 and 4.5%, while other variants such as TPMT*3B and TPMT*2 are absent or very rare [26, 27]. In the present study, the entire gene analysis of TPMT and the most common variations of ITPA in ALL subjects indicated that 8.6% of children are heterozygous for a nonfunctional TPMT allele. The frequency of TPMT*3C in our subjects was 2.22%, which is comparable to studies by Kapoor et al. and Desire et al. that showed variant allele frequencies of 4.5 and 2.6%, respectively [26, 27]. TPMT*2, TPMT*3A, or TPMT*3B variations were absent in our study subjects, which is consistent with a populationbased study by Murugeshan et al. demonstrating the absence of the TPMT*3A variant and very low frequencies of *2 and *3B variants [28]. In subjects of Indian origin residing in the UK and Singapore, very low frequencies of TPMT*3A (0.5%) and TPMT*3C (0.8%) were reported [29, 30]. This could be due to selection bias as the subjects are not representative of a random Indian population. Positive association was observed between the number of high-risk alleles at TPMT/ITPA loci and leucopenia. TPMT*3C subjects had a high frequency of dose adjustments even after having a lower mean dose compared to other genotypes, and 25% experienced relapse or death indicating a high degree of 6-MP-mediated toxicity associated with this variant. TPMT rs2842949 is relatively unexplored, and no data are available to corroborate with leucopenia in other ethnic groups. In the current study, all the variants of this polymorphism showed homozygous mutant phenotype at

Table 5 Epistatic interactions between the variants of thiopurine metabolism influencing hematological toxicity Interacting loci

OR (95% CI)

χ2

P value

ITPA ex2 TPMT*12 × ITPA ex3 ITPA ex2 × TPMT*12 × ITPA ex3 TPMT*3C × ITPA ex2 × TPMT*12 × ITPA ex3

α 4.07 (1.64–10.08) 7.09 (2.59–19.39) 8.33 (3.06–22.65)

8.75 9.66 16.27 19.44

0.003* 0.002*